The composition of bacterial communities is paramount to their function. Intense competitive interactions between bacteria can influence community composition by altering assembly or stability through the production of anti-bacterial molecules. One such ecosystem rife with competitive interactions is the human gastrointestinal microbiota, which harbors an abundance of bacteria from the order Bacteroidales. These bacteria encode the type VI secretion system (T6SS), a contact-dependent toxin delivery pathway shown to mediate potent inter- species competition. Bacteria that encode the T6SS also encode immunity proteins which bind to and specifically inactivate toxins. I have found that Bacteroidales genomes possess arrays of ?orphan immunity? genes in the absence of the T6SS or corresponding toxin. These polyimmunity arrays are found associated with xerD-like (PAX) recombinase genes that suggest an active mechanism of recruitment. I hypothesize that genes within PAX clusters represent selective pressure from past episodes of direct interbacterial antagonism. Therefore, characterization of the function, activity, and biogenesis of PAX clusters offers an avenue for the identification of direct bacterial interactions in vivo and an understanding of their ecological consequences. This proposal aims to augment my interdisciplinary background in cell biology, evolutionary biology, and genetics with new training in germfree mouse experimentation to investigate this new mechanism of adaptation to T6SS-mediated interbacterial antagonism within Bacteroidales and understand its impact on gut microbiome assembly and dynamics. In Aim 1, I propose to characterize the function of genes within PAX clusters through in vitro expression experiments in E. coli and growth competition experiments in vitro and in vivo under conditions that promote contact-dependent interbacterial antagonism. In Aim 2, I will test the hypothesis that the PAX recombinase mediates gene acquisition and excision from PAX clusters via tyrosine recombinase activity in a manner analogous to integrons. In Aim 3, I will extend my characterization of PAX clusters to their use as a new tool for deciphering interbacterial interactions by sequencing new gene insertions after exposure of a PAX-containing strain to different bacterial communities in vivo. I will train with Dr. Lynn Hajjar at the University of Washington Gnotobiotic Animal Core facility to gain experience in using germfree mice to study the role of PAX clusters in the gut microbiota in vivo. Dr. Andrew Goodman at Yale University will collaborate and provide advice and intellectual support. The proposed experiments will lay the foundation for my independent research program, in which bioinformatics, bacterial genetics, and in vivo experimentation will be combined to yield insight into direct interactions within the gut microbiota and the implications of these interactions for human health and disease.